Re: [Tclhttpd-users] question on charsets

[David G. <dav...@po...>]

Shawn Kielty wrote:
> I don't think it follows that it is true in an xml document type declar=
ation
> as well.  Is the document type declaration used for other purposes such=
 as
> validation?

Hey Shawn,

The XML parser reads it so it knows the charset of the bytes that follow.

But, oddly enough with tdom it gets rather difficult because you can't
give [dom parse] the XML declaration because of apparently the strange
way it operates.  tdom instead of externalizing the string to then hand
to expat's XML_Parse() so that the characters you see are the characters
it gets (like any other well behaved extension), subverts the normal
case and hands XML_Parse() Tcl's internal utf-8 rep.

IOW, you get this:
(Desktop) 211 % [dom parse {<?xml version=3D"1.0" encoding=3D"ISO-8859-1"
?><foo>=E4</foo>}] asXML
<foo>=C3=A4</foo>

'=E4' !=3D '=C3=A4' (its internal utf-8 rep)

Even though '=E4' is a member of ISO-8859-1, and I'm claiming the documen=
t
to be as such, is not how tdom works.  I guess one is supposed to
manually read the declaration for a charset, strip the declaration,
convert it with [encoding converfrom], then call [tdom parse].

Whoa..  like that was obvious :(

<bangheadonkeyboard repeat=3D"4" />